diff --git a/README.md b/README.md
index 5b9f764ba7..7cbf3b65e2 100755
--- a/README.md
+++ b/README.md
@@ -23,7 +23,10 @@ from [MLCommons projects](https://mlcommons.org) and [research papers](https://c
 in a unified way on any operating system with any software and hardware
 either natively or inside containers.
 
-Here are some most commonly used examples from the community:
+Here are a few most commonly used examples from the CM users 
+that should run in the same way on Linux, MacOS, Windows and other platforms
+(see [Gettings Started Guide](docs/getting-started.md) to understand 
+how they work and how to reuse them in your projects):
 
 ```bash
 pip install cmind
@@ -31,32 +34,36 @@ pip install cmind
 cm pull repo mlcommons@ck
 
 cm run script "python app image-classification onnx"
+cmr "python app image-classification onnx"
 
-cm run script "download file _wget" --url=https://cKnowledge.org/ai/data/computer_mouse.jpg --verify=no --env.CM_DOWNLOAD_CHECKSUM=45ae5c940233892c2f860efdf0b66e7e
-
-cm run script "python app image-classification onnx" --input=computer_mouse.jpg
+cmr "download file _wget" --url=https://cKnowledge.org/ai/data/computer_mouse.jpg --verify=no --env.CM_DOWNLOAD_CHECKSUM=45ae5c940233892c2f860efdf0b66e7e
 cmr "python app image-classification onnx" --input=computer_mouse.jpg
 
-cm show cache
-cm rm cache -f
-
-cmr "python app image-classification onnx _cuda" --input=computer_mouse.jpg
-
-cmr "cm gui" --script="python app image-classification onnx"
+cm find script "python app image-classification onnx"
 
-cm docker script "python app image-classification onnx" --input=computer_mouse.jpg
-cm docker script "python app image-classification onnx" --input=computer_mouse.jpg -j -docker_it
-cm docker script "python app image-classification onnx" --input=computer_mouse.jpg -j -docker_it
+cmr "get python" --version_min=3.8.0 --name=mlperf-experiments
+cmr "install python-venv" --version_max=3.10.11 --name=mlperf
 
-cmr "get generic-python-lib _package.onnxruntime"
-cmr "get coco dataset _val _2014"
 cmr "get ml-model stable-diffusion"
 cmr "get ml-model huggingface zoo _model-stub.alpindale/Llama-2-13b-ONNX" --model_filename=FP32/LlamaV2_13B_float32.onnx --skip_cache
+cmr "get coco dataset _val _2014"
 
 cm show cache
 cm show cache "get ml-model stable-diffusion"
 
+cmr "get generic-python-lib _package.onnxruntime" --version_min=1.16.0
+cmr "python app image-classification onnx" --input=computer_mouse.jpg
+
 cm rm cache -f
+cmr "python app image-classification onnx" --input=computer_mouse.jpg --adr.onnxruntime.version_max=1.16.0
+
+
+cmr "python app image-classification onnx _cuda" --input=computer_mouse.jpg
+
+cmr "cm gui" --script="python app image-classification onnx"
+
+cm docker script "python app image-classification onnx" --input=computer_mouse.jpg
+cm docker script "python app image-classification onnx" --input=computer_mouse.jpg -j -docker_it
 
 cmr "run common mlperf inference" --implementation=nvidia --model=bert-99 --category=datacenter --division=closed
 cm find script "run common mlperf inference"
@@ -65,8 +72,6 @@ cm pull repo ctuning@cm-reproduce-research-projects
 cmr "reproduce paper micro-2023 victima _install_deps"
 cmr "reproduce paper micro-2023 victima _run" 
 
-...
-
 ```
 
 ```python
@@ -78,8 +83,9 @@ if output['return']==0: print (output)
 ```
 
 
-Collective Mind is a community project being developed by the [MLCommons Task Force on Automation and Reproducibility](https://github.com/mlcommons/ck/blob/master/docs/taskforce.md)
-with great help from [MLCommons (70+ AI organizations)](https://mlcommons.org/),
+Collective Mind is a community project being developed by the 
+[MLCommons Task Force on Automation and Reproducibility](https://github.com/mlcommons/ck/blob/master/docs/taskforce.md)
+with great help from [MLCommons (70+ AI organizations)](https://mlcommons.org/,
 [research community]( https://www.youtube.com/watch?v=7zpeIVwICa4 )
 and [individual contributors](https://github.com/mlcommons/ck/blob/master/CONTRIBUTING.md) -
 we want to have a simple, non-intrusive, technology-agnostic, portable and easily-extensible interface 
@@ -89,7 +95,7 @@ running experiments, processing logs, and reproducing results
 without thinking where and how they run.
 
 That is why we implemented CM as a [small Python library](https://github.com/mlcommons/ck/tree/master/cm) 
-with minimal dependencies (Python 3.7+, git, wget), simple Python API, StreamLit GUI and human-friendly command line.
+with minimal dependencies (Python 3.7+, git, wget), simple Python API, GUI and human-friendly command line.
 CM simply searches for CM scripts by tags or Unique IDs in all pulled Git repositories, automatically generates command lines 
 for a given script or tool on a given platform, updates all paths and environment variables, 
 runs a given automation either natively or inside automatically-generated containers
@@ -115,7 +121,7 @@ or have questions and suggestions.
 
 ### Documentation
 
-* [Getting Started tutorial](docs/getting-started.md)
+* [Getting Started Guide](docs/getting-started.md)
   * [CM interface for MLPerf benchmarks](docs/mlperf)
   * [CM interface for ML and Systems conferences](docs/tutorials/common-interface-to-reproduce-research-projects.md)
   * [CM automation recipes for MLOps and DevOps](cm-mlops/script)
diff --git a/cm-mlops/automation/script/module.py b/cm-mlops/automation/script/module.py
index d67eb0cdec..4f81ece1ad 100644
--- a/cm-mlops/automation/script/module.py
+++ b/cm-mlops/automation/script/module.py
@@ -1704,6 +1704,14 @@ def _update_variation_tags_from_variations(self, variation_tags, variations, var
                 v_static = self._get_name_for_dynamic_variation_tag(v)
                 tmp_variation_tags_static[v_i] = v_static
 
+        combined_variations = [ t for t in variations if ',' in t ]
+        # We support default_variations in the meta of cmbined_variations
+        combined_variations.sort(key=lambda x: x.count(','))
+        ''' By sorting based on the number of variations users can safely override
+            env and state in a larger combined variation
+        '''
+        tmp_combined_variations = {k: False for k in combined_variations}
+
         # Recursively add any base variations specified
         if len(variation_tags) > 0:
             tmp_variations = {k: False for k in variation_tags}
@@ -1748,28 +1756,34 @@ def _update_variation_tags_from_variations(self, variation_tags, variations, var
                             tag_to_append = None
 
                     # default_variations dictionary specifies the default_variation for each variation group. A default variation in a group is turned on if no other variation from that group is turned on and it is not excluded using the '-' prefix
-                    if "default_variations" in variations[variation_name]:
-                        default_base_variations = variations[variation_name]["default_variations"]
-                        for default_base_variation in default_base_variations:
-                            tag_to_append = None
+                    r = self._get_variation_tags_from_default_variations(variations[variation_name], variations, variation_groups, tmp_variation_tags_static, excluded_variation_tags)
+                    if r['return'] > 0:
+                        return r
 
-                            if default_base_variation not in variation_groups:
-                                return {'return': 1, 'error': 'Default variation "{}" is not a valid group. Valid groups are "{}" '.format(default_base_variation, variation_groups)}
+                    variations_to_add = r['variations_to_add']
+                    for t in variations_to_add:
+                        tmp_variations[t] = False
+                        variation_tags.append(t)
 
-                            unique_allowed_variations = variation_groups[default_base_variation]['variations']
-                            # add the default only if none of the variations from the current group is selected and it is not being excluded with - prefix
-                            if len(set(unique_allowed_variations) & set(tmp_variation_tags_static)) == 0 and default_base_variations[default_base_variation] not in excluded_variation_tags and default_base_variations[default_base_variation] not in tmp_variation_tags_static:
-                                tag_to_append = default_base_variations[default_base_variation]
+                    tmp_variations[variation_name] = True
 
-                            if tag_to_append:
-                                if tag_to_append not in variations:
-                                    variation_tag_static = self._get_name_for_dynamic_variation_tag(tag_to_append)
-                                    if not variation_tag_static or variation_tag_static not in variations:
-                                        return {'return': 1, 'error': 'Invalid variation "{}" specified in default variations for the variation "{}" '.format(tag_to_append, variation_name)}
-                                variation_tags.append(tag_to_append)
-                                tmp_variations[tag_to_append] = False
+                    for combined_variation in combined_variations:
+                        if tmp_combined_variations[combined_variation]:
+                            continue
+                        v = combined_variation.split(",")
+                        all_present = set(v).issubset(set(variation_tags))
+                        if all_present:
+                            combined_variation_meta = variations[combined_variation]
+                            tmp_combined_variations[combined_variation] = True
 
-                    tmp_variations[variation_name] = True
+                            r = self._get_variation_tags_from_default_variations(combined_variation_meta, variations, variation_groups, tmp_variation_tags_static, excluded_variation_tags)
+                            if r['return'] > 0:
+                                return r
+
+                            variations_to_add = r['variations_to_add']
+                            for t in variations_to_add:
+                                tmp_variations[t] = False
+                                variation_tags.append(t)
 
                 all_base_processed = True
                 for variation_name in variation_tags:
@@ -1785,6 +1799,31 @@ def _update_variation_tags_from_variations(self, variation_tags, variations, var
         return {'return': 0}
 
 
+    def _get_variation_tags_from_default_variations(self, variation_meta, variations, variation_groups, tmp_variation_tags_static, excluded_variation_tags):
+    # default_variations dictionary specifies the default_variation for each variation group. A default variation in a group is turned on if no other variation from that group is turned on and it is not excluded using the '-' prefix
+
+        tmp_variation_tags = []
+        if "default_variations" in variation_meta:
+            default_base_variations = variation_meta["default_variations"]
+            for default_base_variation in default_base_variations:
+                tag_to_append = None
+
+                if default_base_variation not in variation_groups:
+                    return {'return': 1, 'error': 'Default variation "{}" is not a valid group. Valid groups are "{}" '.format(default_base_variation, variation_groups)}
+
+                unique_allowed_variations = variation_groups[default_base_variation]['variations']
+                # add the default only if none of the variations from the current group is selected and it is not being excluded with - prefix
+                if len(set(unique_allowed_variations) & set(tmp_variation_tags_static)) == 0 and default_base_variations[default_base_variation] not in excluded_variation_tags and default_base_variations[default_base_variation] not in tmp_variation_tags_static:
+                    tag_to_append = default_base_variations[default_base_variation]
+
+                if tag_to_append:
+                    if tag_to_append not in variations:
+                        variation_tag_static = self._get_name_for_dynamic_variation_tag(tag_to_append)
+                        if not variation_tag_static or variation_tag_static not in variations:
+                            return {'return': 1, 'error': 'Invalid variation "{}" specified in default variations for the variation "{}" '.format(tag_to_append, variation_name)}
+                    tmp_variation_tags.append(tag_to_append)
+
+        return {'return': 0, 'variations_to_add': tmp_variation_tags}
 
     ############################################################
     def version(self, i):
@@ -1824,6 +1863,15 @@ def search(self, i):
 
         console = i.get('out') == 'con'
 
+        # Check simplified CMD: cm run script "get compiler"
+        # If artifact has spaces, treat them as tags!
+        artifact = i.get('artifact','')
+        if ' ' in artifact: # or ',' in artifact:
+            del(i['artifact'])
+            if 'parsed_artifact' in i: del(i['parsed_artifact'])
+            # Force substitute tags
+            i['tags']=artifact.replace(' ',',')
+        
         ############################################################################################################
         # Process tags to find script(s) and separate variations 
         # (not needed to find scripts)
@@ -2125,6 +2173,16 @@ def add(self, i):
         tags_list = utils.convert_tags_to_list(i)
         if 'tags' in i: del(i['tags'])
 
+        if len(tags_list)==0:
+            if console:
+                x=input('Please specify a combination of unique tags separated by comma for this script: ')
+                x = x.strip()
+                if x!='':
+                    tags_list = x.split(',')
+
+        if len(tags_list)==0:
+            return {'return':1, 'error':'you must specify a combination of unique tags separate by comman using "--new_tags"'}
+
         # Add placeholder (use common action)
         ii['out']='con'
         ii['common']=True # Avoid recursion - use internal CM add function to add the script artifact
@@ -2150,18 +2208,19 @@ def add(self, i):
 
         # Check if preloaded meta exists
         meta = {
-                 'cache':False,
-                 'new_env_keys':[],
-                 'new_state_keys':[],
-                 'input_mapping':{},
-                 'docker_input_mapping':{},
-                 'deps':[],
-                 'prehook_deps':[],
-                 'posthook_deps':[],
-                 'post_deps':[],
-                 'versions':{},
-                 'variations':{},
-                 'input_description':{}
+                 'cache':False
+# 20240127: Grigori commented that because newly created script meta looks ugly
+#                 'new_env_keys':[],
+#                 'new_state_keys':[],
+#                 'input_mapping':{},
+#                 'docker_input_mapping':{},
+#                 'deps':[],
+#                 'prehook_deps':[],
+#                 'posthook_deps':[],
+#                 'post_deps':[],
+#                 'versions':{},
+#                 'variations':{},
+#                 'input_description':{}
                }
 
         fmeta = os.path.join(template_path, self.cmind.cfg['file_cmeta'])
@@ -2200,6 +2259,7 @@ def add(self, i):
             ii['yaml']=True
 
         ii['automation']='script,5b4e0237da074764'
+
         r_obj=self.cmind.access(ii)
         if r_obj['return']>0: return r_obj
 
diff --git a/cm-mlops/automation/script/module_misc.py b/cm-mlops/automation/script/module_misc.py
index a23502ea6c..a0e55b8408 100644
--- a/cm-mlops/automation/script/module_misc.py
+++ b/cm-mlops/automation/script/module_misc.py
@@ -1384,6 +1384,8 @@ def dockerfile(i):
         fake_run_deps = i.get('fake_run_deps', docker_settings.get('fake_run_deps', False))
         docker_run_final_cmds = docker_settings.get('docker_run_final_cmds', [])
 
+        gh_token = i.get('docker_gh_token')
+
         if i.get('docker_real_run', False):
             fake_run_option = " "
             fake_run_deps = False
@@ -1436,6 +1438,9 @@ def dockerfile(i):
                             'real_run': True
                             }
 
+        if gh_token:
+            cm_docker_input['gh_token'] = gh_token
+
         r = self_module.cmind.access(cm_docker_input)
         if r['return'] > 0:
             return r
@@ -1646,8 +1651,12 @@ def docker(i):
 
         all_gpus = i.get('docker_all_gpus', docker_settings.get('all_gpus'))
 
+        device = i.get('docker_device', docker_settings.get('device'))
+
+        gh_token = i.get('docker_gh_token')
+
+        port_maps = i.get('docker_port_maps', docker_settings.get('port_maps', []))
 
-        
         
 #        # Regenerate run_cmd
 #        if i.get('cmd'):
@@ -1678,7 +1687,6 @@ def docker(i):
         print (run_cmd)
         print ('')
 
-
         cm_docker_input = {'action': 'run',
                            'automation': 'script',
                            'tags': 'run,docker,container',
@@ -1709,6 +1717,15 @@ def docker(i):
         if all_gpus:
             cm_docker_input['all_gpus'] = True
 
+        if device:
+            cm_docker_input['device'] = device
+
+        if gh_token:
+            cm_docker_input['gh_token'] = gh_token
+
+        if port_maps:
+            cm_docker_input['port_maps'] = port_maps
+
         print ('')
 
 
diff --git a/cm-mlops/script/app-image-classification-onnx-py/README-extra.md b/cm-mlops/script/app-image-classification-onnx-py/README-extra.md
index 053a4894fa..e379e2544e 100644
--- a/cm-mlops/script/app-image-classification-onnx-py/README-extra.md
+++ b/cm-mlops/script/app-image-classification-onnx-py/README-extra.md
@@ -4,13 +4,14 @@ See [this tutorial](https://github.com/mlcommons/ck/blob/master/docs/tutorials/m
 
 # Collaborative testing
 
+## Windows 11
+
+* CUDA 11.8; cuDNN 8.7.0; ONNX GPU 1.16.1
+
 ## Windows 10
 
-* ONNX Runtime 1.13.1 with CPU and CUDA
-* CUDA 11.6
-* cuDNN 8.5.0.96
+* CUDA 11.6; cuDNN 8.6.0.96; ONNX GPU 1.13.1
 
 ## Ubuntu 22.04
 
-* ONNX Runtime 1.12.0 with CPU and CUDA
-* CUDA 11.3
+* CUDA 11.3; ONNX 1.12.0
diff --git a/cm-mlops/script/app-image-classification-onnx-py/_cm.yaml b/cm-mlops/script/app-image-classification-onnx-py/_cm.yaml
index 6c07135d7d..c3f869081e 100644
--- a/cm-mlops/script/app-image-classification-onnx-py/_cm.yaml
+++ b/cm-mlops/script/app-image-classification-onnx-py/_cm.yaml
@@ -25,12 +25,20 @@ deps:
   - python
   - python3
   tags: get,python3
+
 - tags: get,cuda
   names:
   - cuda
   enable_if_env:
     USE_CUDA:
     - yes
+- tags: get,cudnn
+  names:
+  - cudnn
+  enable_if_env:
+    USE_CUDA:
+    - yes
+
 - tags: get,dataset,imagenet,image-classification,original
 - tags: get,dataset-aux,imagenet-aux,image-classification
 - tags: get,ml-model,resnet50,_onnx,image-classification
@@ -41,16 +49,20 @@ deps:
 - tags: get,generic-python-lib,_package.numpy
 - tags: get,generic-python-lib,_package.opencv-python
 
+
 - tags: get,generic-python-lib,_onnxruntime
+  names:
+  - onnxruntime
   skip_if_env:
     USE_CUDA:
     - yes
 - tags: get,generic-python-lib,_onnxruntime_gpu
+  names:
+  - onnxruntime
   enable_if_env:
     USE_CUDA:
     - yes
 
-
 variations:
   cuda:
     env:
@@ -61,6 +73,7 @@ input_mapping:
   input: CM_IMAGE
   output: CM_APP_IMAGE_CLASSIFICATION_ONNX_PY_OUTPUT
 
+
 new_env_keys:
   - CM_APP_IMAGE_CLASSIFICATION_ONNX_PY*
 
diff --git a/cm-mlops/script/app-mlperf-inference-cpp/_cm.yaml b/cm-mlops/script/app-mlperf-inference-cpp/_cm.yaml
index 83562e086f..80f28a0ff3 100644
--- a/cm-mlops/script/app-mlperf-inference-cpp/_cm.yaml
+++ b/cm-mlops/script/app-mlperf-inference-cpp/_cm.yaml
@@ -221,6 +221,7 @@ variations:
       CM_MLPERF_LOADGEN_MAX_BATCHSIZE: 8
 
   batch-size.#:
+    group: batch-size
     env:
       CM_MLPERF_LOADGEN_MAX_BATCHSIZE: "#"
 
@@ -242,3 +243,15 @@ variations:
     group: loadgen-scenario
     env:
       CM_MLPERF_LOADGEN_SCENARIO: Server
+
+  multistream,resnet50:
+    default_variations:
+      batch-size: batch-size.8
+
+  offline,resnet50:
+    default_variations:
+      batch-size: batch-size.8
+
+  multistream,retinanet:
+    default_variations:
+      batch-size: batch-size.1
diff --git a/cm-mlops/script/app-mlperf-inference-cpp/src/main.cpp b/cm-mlops/script/app-mlperf-inference-cpp/src/main.cpp
index beb0b72f20..c5a3c809ea 100644
--- a/cm-mlops/script/app-mlperf-inference-cpp/src/main.cpp
+++ b/cm-mlops/script/app-mlperf-inference-cpp/src/main.cpp
@@ -42,7 +42,8 @@ class InputSettings {
         if (mode_name == "performance")
             mode_name = "PerformanceOnly";
         query_count_override = std::stol(getenv("CM_MLPERF_LOADGEN_QUERY_COUNT", "0"));
-        performance_sample_count = std::stol(getenv("CM_MLPERF_LOADGEN_PERFORMANCE_SAMPLE_COUNT", "1024"));
+	query_count_override = 0;
+        performance_sample_count = std::stol(getenv("CM_MLPERF_LOADGEN_PERFORMANCE_SAMPLE_COUNT", "0"));
         batch_size = std::stol(getenv("CM_MLPERF_LOADGEN_MAX_BATCHSIZE", "32"));
         std::cout << "MLPerf Conf path: " << mlperf_conf_path << std::endl;
         std::cout << "User Conf path: " << user_conf_path << std::endl;
@@ -52,6 +53,7 @@ class InputSettings {
         std::cout << "Mode: " << mode_name << std::endl;
         std::cout << "Batch size: " << batch_size << std::endl;
         std::cout << "Query count override: " << query_count_override << std::endl;
+        std::cout << "Performance sample count override in application: " << performance_sample_count << std::endl;
     }
 
     std::string mlperf_conf_path;
diff --git a/cm-mlops/script/app-mlperf-inference-reference/_cm.yaml b/cm-mlops/script/app-mlperf-inference-reference/_cm.yaml
index a0f0690ae5..650bdffaff 100644
--- a/cm-mlops/script/app-mlperf-inference-reference/_cm.yaml
+++ b/cm-mlops/script/app-mlperf-inference-reference/_cm.yaml
@@ -840,6 +840,7 @@ variations:
     group: models
     env:
       CM_MODEL: stable-diffusion-xl
+      CM_NUM_THREADS: "1"
     deps:
       - tags: get,generic-python-lib,_package.diffusers
         names:
diff --git a/cm-mlops/script/app-mlperf-inference/_cm.yaml b/cm-mlops/script/app-mlperf-inference/_cm.yaml
index 118b521b1a..e517c90d21 100644
--- a/cm-mlops/script/app-mlperf-inference/_cm.yaml
+++ b/cm-mlops/script/app-mlperf-inference/_cm.yaml
@@ -131,6 +131,8 @@ variations:
           CM_SKIP_RUN:
             - yes
 
+  mil:
+    alias: cpp
 
   tflite-cpp:
     default_variations:
@@ -383,6 +385,10 @@ variations:
     env:
       CM_MODEL:
         stable-diffusion-xl
+    default_variations:
+      precision: float16
+      backend: pytorch
+      device: cuda
     add_deps_recursive:
       mlperf-inference-implementation:
         tags: _sdxl
diff --git a/cm-mlops/script/app-mlperf-inference/customize.py b/cm-mlops/script/app-mlperf-inference/customize.py
index bd5447cac9..21b95b67d2 100644
--- a/cm-mlops/script/app-mlperf-inference/customize.py
+++ b/cm-mlops/script/app-mlperf-inference/customize.py
@@ -90,8 +90,15 @@ def postprocess(i):
         accuracy_log_file_option_name = " --log_file "
         datatype_option = " --output_dtype "+env['CM_SQUAD_ACCURACY_DTYPE']
 
+      elif 'stable-diffusion-xl' in model:
+          pass #No compliance check for now
+      elif 'gpt' in model:
+          pass #No compliance check for now
+      elif 'llama2-70b' in model:
+          pass #No compliance check for now
       else:
-        return {'return': 1, 'error': f'Accuracy paths not done for model {model}'}
+        pass # Not giving an error now. But accuracy paths need to be done for other benchmarks which may need the non-determinism test
+        #return {'return': 1, 'error': f'Accuracy paths not done for model {model}'}
     scenario = env['CM_MLPERF_LOADGEN_SCENARIO']
 
     #if env.get("CM_MLPERF_FIND_PERFORMANCE_MODE", '') == "yes" and mode == "performance" and scenario != "Server":
diff --git a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/_cm.yaml b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/_cm.yaml
index f42b4067c6..020eadfec4 100644
--- a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/_cm.yaml
+++ b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/_cm.yaml
@@ -5,10 +5,19 @@ cache: false
 category: MLPerf benchmark support
 deps:
   - tags: detect,cpu
-
+default_env:
+  DIVISION: open
+  CATEGORY: edge
 docker_input_mapping: {}
 input_description: {}
-input_mapping: {}
+input_mapping:
+  models: MODELS
+  backends: BACKENDS
+  devices: DEVICES
+  division: DIVISION
+  category: CATEGORY
+  power_server: POWER_SERVER
+  power_server_port: POWER_SERVER_PORT
 new_env_keys: []
 new_state_keys: []
 post_deps: []
@@ -105,7 +114,7 @@ variations:
     env:
       POWER: yes
     default_env:
-      POWER_SERVER_IP: 192.168.0.15
+      POWER_SERVER: 192.168.0.15
       POWER_SERVER_PORT: 4950
 
 
@@ -141,5 +150,3 @@ variations:
     env:
       CATEGORY: edge
       DIVISION: closed
-
-versions: {}
diff --git a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/customize.py b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/customize.py
index 2902feffb4..b62daab258 100644
--- a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/customize.py
+++ b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/customize.py
@@ -29,7 +29,7 @@ def preprocess(i):
     power = env.get('POWER', '')
 
     if str(power).lower() in [ "yes", "true" ]:
-        POWER_STRING = " --power yes --adr.mlperf-power-client.power_server=" + env.get('POWER_SERVER', '192.168.0.15') + " --adr.mlperf-power-client.port=" + env.get('POWER_SERVER_PORT', '4950') + " "
+        POWER_STRING = " --power=yes --adr.mlperf-power-client.power_server=" + env.get('POWER_SERVER', '192.168.0.15') + " --adr.mlperf-power-client.port=" + str(env.get('POWER_SERVER_PORT', '4950')) + " "
     else:
         POWER_STRING = ""
 
@@ -37,7 +37,12 @@ def preprocess(i):
         return {'return': 1, 'error': 'No device specified. Please set one or more (comma separated) of {cpu, qaic, cuda, rocm} for --env.DEVICES=<>'}
 
     cmds = []
-    run_script_content = '#!/bin/bash\nsource '+ os.path.join(script_path, "run-template.sh")
+    run_script_content = '#!/bin/bash\n\n'
+    run_script_content += "POWER_STRING=\"" +POWER_STRING +"\"\n"
+    run_script_content += "DIVISION=\"" + env['DIVISION'] +"\"\n"
+    run_script_content += "CATEGORY=\"" + env['CATEGORY'] +"\"\n"
+    run_script_content += 'source '+ os.path.join(script_path, "run-template.sh") + "\nPOWER_STRING=\"" +POWER_STRING +"\"\n\n"
+
     run_file_name = 'tmp-'+implementation+'-run'
 
     for model in models:
@@ -107,7 +112,7 @@ def preprocess(i):
                             test_query_count = 10000
                         else:
                             test_query_count = 1000
-                    cmd = f'run_test "{model} "{backend}" "{test_query_count}" "{implementation}" "{device}" "$find_performance_cmd"'
+                    cmd = f'run_test "{model}" "{backend}" "{test_query_count}" "{implementation}" "{device}" "$find_performance_cmd"'
                     cmds.append(cmd)
                     #second argument is unused for submission_cmd
                 cmd = f'run_test "{model}" "{backend}" "100" "{implementation}" "{device}" "$submission_cmd"'
@@ -118,6 +123,10 @@ def preprocess(i):
         f.write(run_script_content)
     print(run_script_content)
 
+    run_script_input = i['run_script_input']
+    r = automation.run_native_script({'run_script_input':run_script_input, 'env':env, 'script_name':run_file_name})
+
+    if r['return']>0: return r
     
     return {'return':0}
 
diff --git a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/run-template.sh b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/run-template.sh
index 0f75941f18..c396c42e4d 100644
--- a/cm-mlops/script/benchmark-any-mlperf-inference-implementation/run-template.sh
+++ b/cm-mlops/script/benchmark-any-mlperf-inference-implementation/run-template.sh
@@ -29,16 +29,15 @@ category=$CATEGORY
 rerun=$RERUN
 
 function run_test() {
-  backend=$1
-  test_query_count=$2
-  implementation=$3
-  device=$4
-  model=$5
+  model=$1
+  backend=$2
+  test_query_count=$3
+  implementation=$4
+  device=$5
   run "$6"
 }
 
 #power=' --power=yes --adr.mlperf-power-client.power_server=192.168.0.15 --adr.mlperf-power-client.port=4950 '
-power=${POWER_STRING}
 results_dir=$HOME/results_dir
 
 #Add your run commands here...
@@ -49,22 +48,22 @@ find_performance_cmd='cm run script --tags=generate-run-cmds,inference,_find-per
 submission_cmd='cm run script --tags=generate-run-cmds,inference,_submission,_all-scenarios \
 --model=$model --implementation=$implementation --device=$device --backend=$backend \
 --category=$category --division=$division  --quiet --results_dir=$results_dir \
---skip_submission_generation=yes --execution-mode=valid $power'
+--skip_submission_generation=yes --execution-mode=valid ${POWER_STRING}'
 
 submission_cmd_scenario='cm run script --tags=generate-run-cmds,inference,_submission  --scenario=$scenario \
 --model=$model --implementation=$implementation --device=$device --backend=$backend \
 --category=$category --division=$division  --quiet --results_dir=$results_dir \
---skip_submission_generation=yes --execution-mode=valid $power'
+--skip_submission_generation=yes --execution-mode=valid ${POWER_STRING}'
 
 readme_cmd_single='cm run script --tags=generate-run-cmds,inference,_populate-readme --scenario=$scenario \
 --model=$model --implementation=$implementation --device=$device --backend=$backend \
 --category=$category --division=$division  --quiet --results_dir=$results_dir \
---skip_submission_generation=yes --execution-mode=valid $power'
+--skip_submission_generation=yes --execution-mode=valid ${POWER_STRING}'
 
 readme_cmd='cm run script --tags=generate-run-cmds,inference,_populate-readme,_all-scenarios \
 --model=$model --implementation=$implementation --device=$device --backend=$backend \
 --category=$category --division=$division  --quiet --results_dir=$results_dir \
---skip_submission_generation=yes --execution-mode=valid $power'
+--skip_submission_generation=yes --execution-mode=valid ${POWER_STRING}'
 
 
 tflite_accuracy_cmd='cm run script --tags=run,mobilenet-models,_tflite,_accuracy-only$extra_tags \
@@ -73,13 +72,13 @@ ${extra_option} \
 --results_dir=$results_dir'
 
 tflite_performance_cmd='cm run script --tags=run,mobilenet-models,_tflite,_performance-only$extra_tags \
-${POWER} \
+${POWER_STRING} \
 --adr.compiler.tags=gcc \
 ${extra_option} \
 --results_dir=$results_dir'
 
 tflite_readme_cmd='cm run script --tags=run,mobilenet-models,_tflite,_populate-readme$extra_tags \
-${POWER} \
+${POWER_STRING} \
 --adr.compiler.tags=gcc \
 ${extra_option} \
 --results_dir=$results_dir'
diff --git a/cm-mlops/script/build-dockerfile/customize.py b/cm-mlops/script/build-dockerfile/customize.py
index c5c473d8de..15d52fe1c1 100644
--- a/cm-mlops/script/build-dockerfile/customize.py
+++ b/cm-mlops/script/build-dockerfile/customize.py
@@ -174,7 +174,6 @@ def preprocess(i):
     if gh_token:
         run_cmd_extra = " --env.CM_GH_TOKEN=$CM_GH_TOKEN"
 
-
     f.write(EOL+'# Run commands' + EOL)
     for comment in env.get('CM_DOCKER_RUN_COMMENTS', []):
         f.write(comment + EOL)
diff --git a/cm-mlops/script/build-dockerfile/dockerinfo.json b/cm-mlops/script/build-dockerfile/dockerinfo.json
index 12a654d617..5fba91b03f 100644
--- a/cm-mlops/script/build-dockerfile/dockerinfo.json
+++ b/cm-mlops/script/build-dockerfile/dockerinfo.json
@@ -1,6 +1,6 @@
 {
   "python-packages": [
-    "cmind", "requests"
+    "cmind", "requests", "giturlparse"
   ],
   "ARGS": [
     "CM_GH_TOKEN"
diff --git a/cm-mlops/script/create-patch/_cm.yaml b/cm-mlops/script/create-patch/_cm.yaml
new file mode 100644
index 0000000000..fa6a3bd6c1
--- /dev/null
+++ b/cm-mlops/script/create-patch/_cm.yaml
@@ -0,0 +1,16 @@
+uid: 0659dc1f75664c65
+alias: create-patch
+
+automation_alias: script
+automation_uid: 5b4e0237da074764
+
+cache: false
+
+input_mapping:
+  new: CM_CREATE_PATCH_NEW
+  old: CM_CREATE_PATCH_OLD
+
+tags:
+- create
+- patch
+
diff --git a/cm-mlops/script/create-patch/customize.py b/cm-mlops/script/create-patch/customize.py
new file mode 100644
index 0000000000..f716d8303d
--- /dev/null
+++ b/cm-mlops/script/create-patch/customize.py
@@ -0,0 +1,43 @@
+from cmind import utils
+import os
+
+def preprocess(i):
+
+    os_info = i['os_info']
+
+    env = i['env']
+
+    meta = i['meta']
+
+    automation = i['automation']
+
+    quiet = (env.get('CM_QUIET', False) == 'yes')
+
+
+    new_dir = env.get('CM_CREATE_PATCH_NEW', '')
+    if new_dir == '':
+        return {'return':1, 'error':'specify NEW directory using --new'}
+    if not os.path.isdir(new_dir):
+        return {'return':1, 'error':'NEW directory doesn\'t exist {}'.format(new_dir)}
+
+    old_dir = env.get('CM_CREATE_PATCH_OLD', '')
+    if old_dir == '':
+        return {'return':1, 'error':'specify OLD directory using --new'}
+    if not os.path.isdir(old_dir):
+        return {'return':1, 'error':'OLD directory doesn\'t exist {}'.format(old_dir)}
+
+    cmd = 'diff -Naur {} {} > patch.patch'.format(old_dir, new_dir)
+
+    print ('Running command:')
+    print (cmd)
+
+    os.system(cmd)
+
+
+    return {'return':0}
+
+def postprocess(i):
+
+    env = i['env']
+
+    return {'return':0}
diff --git a/cm-mlops/script/detect-sudo/_cm.yaml b/cm-mlops/script/detect-sudo/_cm.yaml
index 5f8da3f890..56c83b8304 100644
--- a/cm-mlops/script/detect-sudo/_cm.yaml
+++ b/cm-mlops/script/detect-sudo/_cm.yaml
@@ -1,22 +1,17 @@
+uid: 1d47ffc556e248dc
 alias: detect-sudo
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: false
+
 category: DevOps automation
-deps: []
-docker_input_mapping: {}
-input_description: {}
-input_mapping: {}
+
 new_env_keys:
   - CM_SUDO_*
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
+
 tags:
 - detect
 - sudo
 - access
-uid: 1d47ffc556e248dc
-variations: {}
-versions: {}
diff --git a/cm-mlops/script/fail/README-extra.md b/cm-mlops/script/fail/README-extra.md
new file mode 100644
index 0000000000..582991f6d2
--- /dev/null
+++ b/cm-mlops/script/fail/README-extra.md
@@ -0,0 +1 @@
+# CM script
diff --git a/cm-mlops/script/fail/_cm.yaml b/cm-mlops/script/fail/_cm.yaml
new file mode 100644
index 0000000000..4038d83b5b
--- /dev/null
+++ b/cm-mlops/script/fail/_cm.yaml
@@ -0,0 +1,16 @@
+uid: 3aaee82e19d243cd
+alias: fail
+
+automation_alias: script
+automation_uid: 5b4e0237da074764
+
+cache: false
+
+tags:
+- fail
+- filter
+
+variations:
+  windows:
+    env:
+      CM_FAIL_WINDOWS: true
diff --git a/cm-mlops/script/fail/customize.py b/cm-mlops/script/fail/customize.py
new file mode 100644
index 0000000000..855c39b5bb
--- /dev/null
+++ b/cm-mlops/script/fail/customize.py
@@ -0,0 +1,28 @@
+from cmind import utils
+import os
+
+def preprocess(i):
+
+    os_info = i['os_info']
+
+    env = i['env']
+
+    meta = i['meta']
+
+    automation = i['automation']
+
+    quiet = (env.get('CM_QUIET', False) == 'yes')
+
+    # Checking conditions
+    if env.get('CM_FAIL_WINDOWS','').lower()=='true':
+        if os_info['platform'] == 'windows':
+            return {'return':1, 'error': 'CM detected fail condition: running on Windows'}
+
+
+    return {'return':0}
+
+def postprocess(i):
+
+    env = i['env']
+
+    return {'return':0}
diff --git a/cm-mlops/script/get-compiler-rust/_cm.yaml b/cm-mlops/script/get-compiler-rust/_cm.yaml
index c0a7d010b0..ca1a372d9c 100644
--- a/cm-mlops/script/get-compiler-rust/_cm.yaml
+++ b/cm-mlops/script/get-compiler-rust/_cm.yaml
@@ -1,23 +1,19 @@
+uid: 97ffbd9e537b4b59
 alias: get-compiler-rust
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: true
+
 category: Compiler automation
+
 deps:
  - tags: get,python3
 
-docker_input_mapping: {}
-input_description: {}
-input_mapping: {}
 new_env_keys:
 - "+PATH"
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
+
 tags:
 - get
 - rust-compiler
-uid: 97ffbd9e537b4b59
-variations: {}
-versions: {}
diff --git a/cm-mlops/script/get-cuda-devices/_cm.json b/cm-mlops/script/get-cuda-devices/_cm.json
index ffa5ede93d..2a01c191df 100644
--- a/cm-mlops/script/get-cuda-devices/_cm.json
+++ b/cm-mlops/script/get-cuda-devices/_cm.json
@@ -2,6 +2,8 @@
   "alias": "get-cuda-devices",
   "automation_alias": "script",
   "automation_uid": "5b4e0237da074764",
+  "can_force_cache": true,
+  "cache": false,
   "category": "CUDA automation",
   "clean_files": [
     "tmp-run.out"
diff --git a/cm-mlops/script/get-cuda-devices/run.sh b/cm-mlops/script/get-cuda-devices/run.sh
index 87c51fbb77..9c774758fc 100644
--- a/cm-mlops/script/get-cuda-devices/run.sh
+++ b/cm-mlops/script/get-cuda-devices/run.sh
@@ -25,7 +25,8 @@ echo ""
 echo "Running program ..."
 echo ""
 
-cd ${CM_TMP_CURRENT_PATH}
+cd -
+#${CM_TMP_CURRENT_PATH}
 
 ${CM_TMP_CURRENT_SCRIPT_PATH}/a.out > tmp-run.out
 test $? -eq 0 || exit 1
diff --git a/cm-mlops/script/get-cuda/_cm.json b/cm-mlops/script/get-cuda/_cm.json
index 65e2956ecb..d397624bd9 100644
--- a/cm-mlops/script/get-cuda/_cm.json
+++ b/cm-mlops/script/get-cuda/_cm.json
@@ -13,7 +13,8 @@
     "run": false
   },
   "input_mapping": {
-    "cudnn_tar_path": "CM_CUDNN_TAR_FILE_PATH"
+    "cudnn_tar_path": "CM_CUDNN_TAR_FILE_PATH",
+    "cudnn_tar_file": "CM_CUDNN_TAR_FILE_PATH"
   },
   "new_env_keys": [
     "CUDA_HOME",
diff --git a/cm-mlops/script/get-cudnn/README-extra.md b/cm-mlops/script/get-cudnn/README-extra.md
new file mode 100644
index 0000000000..3746808138
--- /dev/null
+++ b/cm-mlops/script/get-cudnn/README-extra.md
@@ -0,0 +1,3 @@
+# TBD
+
+We need to add detection of cuDNN version on Windows, Linux and MacOS
diff --git a/cm-mlops/script/get-cudnn/customize.py b/cm-mlops/script/get-cudnn/customize.py
index fbc4e67077..6d57c5ac32 100644
--- a/cm-mlops/script/get-cudnn/customize.py
+++ b/cm-mlops/script/get-cudnn/customize.py
@@ -28,6 +28,7 @@ def preprocess(i):
 
        libfilename = extra_pre + 'cudnn.' +extra_ext
        env['CM_CUDNN_VERSION'] = 'vdetected'
+
        if os.path.exists(os.path.join(cuda_path_lib, libfilename)):
            env['CM_CUDA_PATH_LIB_CUDNN'] = env['CM_CUDA_PATH_LIB']
            return {'return': 0}
diff --git a/cm-mlops/script/get-dataset-imagenet-calibration/_cm.yaml b/cm-mlops/script/get-dataset-imagenet-calibration/_cm.yaml
index 7d973c68a3..212c506a1f 100644
--- a/cm-mlops/script/get-dataset-imagenet-calibration/_cm.yaml
+++ b/cm-mlops/script/get-dataset-imagenet-calibration/_cm.yaml
@@ -1,7 +1,11 @@
+uid: 30361fad3dff49ff
 alias: get-dataset-imagenet-calibration
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: true
+
 deps:
   - tags: download,file
     force_cache: true
@@ -11,21 +15,15 @@ deps:
     env:
         CM_DOWNLOAD_FINAL_ENV_NAME: CM_MLPERF_IMAGENET_CALIBRATION_LIST_FILE_WITH_PATH
 
-docker_input_mapping: {}
-input_description: {}
-input_mapping: {}
 new_env_keys:
 - CM_MLPERF_IMAGENET_CALIBRATION_LIST_FILE_WITH_PATH
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
+
 tags:
 - get
 - dataset
 - imagenet
 - calibration
-uid: 30361fad3dff49ff
+
 variations:
   mlperf.option1:
     group: calibration-option
diff --git a/cm-mlops/script/get-dataset-openimages-calibration/_cm.yaml b/cm-mlops/script/get-dataset-openimages-calibration/_cm.yaml
index 2ae594f52b..c2ffadffcd 100644
--- a/cm-mlops/script/get-dataset-openimages-calibration/_cm.yaml
+++ b/cm-mlops/script/get-dataset-openimages-calibration/_cm.yaml
@@ -1,7 +1,11 @@
+uid: 27228976bb084dd0
 alias: get-dataset-openimages-calibration
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: true
+
 deps:
   - tags: download,file
     force_cache: true
@@ -11,21 +15,15 @@ deps:
     env:
         CM_DOWNLOAD_FINAL_ENV_NAME: CM_MLPERF_OPENIMAGES_CALIBRATION_LIST_FILE_WITH_PATH
 
-docker_input_mapping: {}
-input_description: {}
-input_mapping: {}
 new_env_keys:
 - CM_MLPERF_OPENIMAGES_CALIBRATION_LIST_FILE_WITH_PATH
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
+
 tags:
 - get
 - dataset
 - openimages
 - calibration
-uid: 27228976bb084dd0
+
 variations:
   mlperf.option1:
     group: calibration-option
diff --git a/cm-mlops/script/get-generic-python-lib/customize.py b/cm-mlops/script/get-generic-python-lib/customize.py
index 1014fe2e05..5027d82376 100644
--- a/cm-mlops/script/get-generic-python-lib/customize.py
+++ b/cm-mlops/script/get-generic-python-lib/customize.py
@@ -15,6 +15,7 @@ def preprocess(i):
         return automation._available_variations({'meta':meta})
 
     if package_name == "onnxruntime_gpu":
+        # https://onnxruntime.ai/docs/execution-providers/CUDA-ExecutionProvider.html#requirements
         cuda_version_split = env['CM_CUDA_VERSION'].split(".")
         if int(cuda_version_split[0]) >= 12:
             # env['CM_INSTALL_ONNXRUNTIME_GPU_FROM_SRC'] = "yes"
diff --git a/cm-mlops/script/get-ipol-src/README.extra.md b/cm-mlops/script/get-ipol-src/README.extra.md
new file mode 100644
index 0000000000..1618d0ed0b
--- /dev/null
+++ b/cm-mlops/script/get-ipol-src/README.extra.md
@@ -0,0 +1 @@
+20240127: Grigori added patch to support latest PIL
diff --git a/cm-mlops/script/get-ipol-src/customize.py b/cm-mlops/script/get-ipol-src/customize.py
index b4da451adf..f2b5dd1fca 100644
--- a/cm-mlops/script/get-ipol-src/customize.py
+++ b/cm-mlops/script/get-ipol-src/customize.py
@@ -7,6 +7,8 @@ def preprocess(i):
 
     env = i['env']
 
+    script_path = i['artifact'].path
+
     automation = i['automation']
 
     cm = automation.cmind
@@ -47,4 +49,10 @@ def preprocess(i):
 
     env['CM_IPOL_PATH']=os.path.join(path, subdir)
 
+    # Applying patch
+    cmd = 'patch -p0 < {}'.format(os.path.join(script_path, 'patch', '20240127.patch'))
+
+    print ('Patching code: {}'.format(cmd))
+    os.system(cmd)
+
     return {'return':0}
diff --git a/cm-mlops/script/get-ipol-src/patch/20240127.patch b/cm-mlops/script/get-ipol-src/patch/20240127.patch
new file mode 100644
index 0000000000..6610d0ceea
--- /dev/null
+++ b/cm-mlops/script/get-ipol-src/patch/20240127.patch
@@ -0,0 +1,10 @@
+diff -Naur 439-master/main.py 439-master.new/main.py
+--- 439-master/main.py	Sat Jan 27 22:11:55 2024
++++ 439-master.new/main.py	Sat Jan 27 22:06:51 2024
+@@ -135,5 +135,5 @@
+     args = parser.parse_args()
+     #print('before plume detection', os.path.dirname(os.path.realpath('__file__')), file=sys.stderr)
+     p = compute_map(args.input_0,args.input_1)
+-    imageio.imsave("cm.png", ((255*p[0,:,:])).numpy())
++    imageio.imsave("cm.png", np.array((255*p[0,:,:]).numpy(), np.uint8))
+ 
diff --git a/cm-mlops/script/get-mlperf-inference-src/_cm.json b/cm-mlops/script/get-mlperf-inference-src/_cm.json
index 8db3b73b47..6b0dbb6bfd 100644
--- a/cm-mlops/script/get-mlperf-inference-src/_cm.json
+++ b/cm-mlops/script/get-mlperf-inference-src/_cm.json
@@ -165,7 +165,7 @@
   "versions": {
     "master": {
       "env": {
-        "CM_MLPERF_LAST_RELEASE": "v3.1",
+        "CM_MLPERF_LAST_RELEASE": "v4.0",
         "CM_GIT_CHECKOUT": "master"
       }
     },
diff --git a/cm-mlops/script/get-preprocessed-dataset-squad/_cm.yaml b/cm-mlops/script/get-preprocessed-dataset-squad/_cm.yaml
index a428b6f081..cff348c266 100644
--- a/cm-mlops/script/get-preprocessed-dataset-squad/_cm.yaml
+++ b/cm-mlops/script/get-preprocessed-dataset-squad/_cm.yaml
@@ -1,8 +1,13 @@
+uid: 7cd1d9b7e8af4788
 alias: get-preprocessed-dataset-squad
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: true
+
 category: AI/ML datasets
+
 deps:
   - tags: get,python3
     names:
@@ -21,25 +26,19 @@ deps:
   - tags: get,generic-python-lib,_package.transformers
   - tags: get,generic-python-lib,_package.tensorflow
 
-docker_input_mapping: {}
-input_description: {}
 env:
   CM_DATASET_MAX_QUERY_LENGTH: 64
 
-input_mapping: {}
 new_env_keys:
   - CM_DATASET_SQUAD_TOKENIZED_*
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
+
 tags:
 - get
 - dataset
 - preprocessed
 - tokenized
 - squad
-uid: 7cd1d9b7e8af4788
+
 variations:
   calib1:
     group: calibration-set
@@ -92,5 +91,3 @@ variations:
         inherit_variation_tags: true
         skip_inherit_variation_groups:
         - packing
-
-versions: {}
diff --git a/cm-mlops/script/get-spec-ptd/_cm.json b/cm-mlops/script/get-spec-ptd/_cm.json
index ccbffa3838..2e7a2c1cb8 100644
--- a/cm-mlops/script/get-spec-ptd/_cm.json
+++ b/cm-mlops/script/get-spec-ptd/_cm.json
@@ -8,7 +8,6 @@
     "CM_GIT_CHECKOUT": "main",
     "CM_GIT_DEPTH": "--depth 1",
     "CM_GIT_PATCH": "no",
-    "CM_GIT_AUTH": "yes",
     "CM_GIT_RECURSE_SUBMODULES": " ",
     "CM_GIT_URL": "https://github.com/mlcommons/power.git"
   },
@@ -33,7 +32,11 @@
     {
       "tags": "get,git,repo,_repo.https://github.com/mlcommons/power",
       "extra_cache_tags": "mlperf,power,spec,ptdaemon,ptd",
+      "force_env_keys": [
+        "CM_GIT_*"
+      ],
       "env": {
+        "CM_GIT_AUTH": "yes",
         "CM_GIT_CHECKOUT_PATH_ENV_NAME": "CM_MLPERF_POWER_SOURCE"
       }
     }
diff --git a/cm-mlops/script/get-spec-ptd/customize.py b/cm-mlops/script/get-spec-ptd/customize.py
index 7cab98136a..250ddd887b 100644
--- a/cm-mlops/script/get-spec-ptd/customize.py
+++ b/cm-mlops/script/get-spec-ptd/customize.py
@@ -14,8 +14,12 @@ def postprocess(i):
     env = i['env']
     state = i['state']
 
+    if env['CM_HOST_OS_TYPE'].lower() == "windows":
+        binary_name = "ptd-windows-x86.exe"
+    else:
+        binary_name = "ptd-linux-x86"
     if 'CM_MLPERF_PTD_PATH' not in env:
-        env['CM_MLPERF_PTD_PATH'] = os.path.join(env['CM_MLPERF_POWER_SOURCE'], 'inference_v1.0', 'ptd-linux-x86')
+        env['CM_MLPERF_PTD_PATH'] = os.path.join(env['CM_MLPERF_POWER_SOURCE'], 'inference_v1.0', binary_name)
     env['CM_SPEC_PTD_PATH'] = env['CM_MLPERF_PTD_PATH']
 
     return {'return':0}
diff --git a/cm-mlops/script/install-onnxruntime-from-src/_cm.json b/cm-mlops/script/install-onnxruntime-from-src/_cm.json
index 9b16611d33..a06b735852 100644
--- a/cm-mlops/script/install-onnxruntime-from-src/_cm.json
+++ b/cm-mlops/script/install-onnxruntime-from-src/_cm.json
@@ -11,6 +11,9 @@
     {
       "tags": "detect,cpu"
     },
+    {
+      "tags": "fail,filter,_windows"
+    },
     {
       "names": [
         "python",
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/README-extra.md b/cm-mlops/script/reproduce-ipol-paper-2022-439/README-extra.md
index 2c5d45c526..fa37618881 100644
--- a/cm-mlops/script/reproduce-ipol-paper-2022-439/README-extra.md
+++ b/cm-mlops/script/reproduce-ipol-paper-2022-439/README-extra.md
@@ -49,12 +49,15 @@ cm run script "get ipol src" --year=2022 --number=439
 cm show cache --tags=ipol,src
 ```
 
-4. Run demo (CM will detect or install missing dependencies)
+4. Download sample images and run demo (CM will detect or install missing dependencies)
 ```bash
+cm run script "download file _wget" --url=https://cKnowledge.org/ai/data/ipol-paper-2024-439-sample-image-1.png --verify=no --env.CM_DOWNLOAD_CHECKSUM=850639287ad23194576582680c2ecfc3
+cm run script "download file _wget" --url=https://cKnowledge.org/ai/data/ipol-paper-2024-439-sample-image-2.png --verify=no --env.CM_DOWNLOAD_CHECKSUM=31364c03d91873ed2d244cce6d664dd0
 cm run script "reproduce ipol 2022-439"
+cm run script "reproduce ipol 2022-439" --adr.torch.version=1.13.1 --adr.torchvision.version=0.14.1
 ```
 
-This script will use these [sample images](https://github.com/mlcommons/ck/tree/master/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images)
+This script will use 2 sample images from this paper
 and should produce *diff.png* in the current directory.
 
 ## Usage with different images
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/_cm.yaml b/cm-mlops/script/reproduce-ipol-paper-2022-439/_cm.yaml
index cf08b1f4dc..bd7c9e1401 100644
--- a/cm-mlops/script/reproduce-ipol-paper-2022-439/_cm.yaml
+++ b/cm-mlops/script/reproduce-ipol-paper-2022-439/_cm.yaml
@@ -22,6 +22,9 @@ deps:
 - tags: get,generic-python-lib,_torch
   names:
   - torch
+- tags: get,generic-python-lib,_torchvision
+  names:
+  - torchvision
 
 tags:
 - app
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/customize.py b/cm-mlops/script/reproduce-ipol-paper-2022-439/customize.py
index 1409115a1f..6b57ab932f 100644
--- a/cm-mlops/script/reproduce-ipol-paper-2022-439/customize.py
+++ b/cm-mlops/script/reproduce-ipol-paper-2022-439/customize.py
@@ -6,8 +6,29 @@ def preprocess(i):
 
     env = i['env']
 
+    # Check if input files are empty and add files
+    input_file_1 = env.get('CM_INPUT_1','')
+    if input_file_1 == '': input_file_1 = 'ipol-paper-2024-439-sample-image-1.png'
+
+    if not os.path.isfile(input_file_1):
+        return {'return':1, 'error':'input file 1 "{}" not found'.format(input_file_1)}
+
+    env['CM_INPUT_1']=os.path.abspath(input_file_1)
+
+    input_file_2 = env.get('CM_INPUT_2','')
+    if input_file_2 == '': input_file_2 = 'ipol-paper-2024-439-sample-image-2.png'
+
+    if not os.path.isfile(input_file_2):
+        return {'return':1, 'error':'input file 2 "{}" not found'.format(input_file_2)}
+
+    env['CM_INPUT_2']=os.path.abspath(input_file_2)
+
     return {'return':0}
 
 def postprocess(i):
 
+    print ('')
+    print ('Please check "diff.png"')
+    print ('')
+
     return {'return':0}
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/requirements.txt b/cm-mlops/script/reproduce-ipol-paper-2022-439/requirements.txt
index 0778518560..82a4d6034b 100644
--- a/cm-mlops/script/reproduce-ipol-paper-2022-439/requirements.txt
+++ b/cm-mlops/script/reproduce-ipol-paper-2022-439/requirements.txt
@@ -1,7 +1,5 @@
 jupyter
 numpy
 imageio
-torch
-torchvision
 IPython
 scikit-image
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/run.bat b/cm-mlops/script/reproduce-ipol-paper-2022-439/run.bat
index bd1506709e..7aafa4a34a 100644
--- a/cm-mlops/script/reproduce-ipol-paper-2022-439/run.bat
+++ b/cm-mlops/script/reproduce-ipol-paper-2022-439/run.bat
@@ -17,15 +17,6 @@ cd %CM_IPOL_PATH%
 
 echo Current path in CM cache: %cd%
 
-rem Check default images
-if "%CM_INPUT_1%" == "" (
-  set CM_INPUT_1=%CM_TMP_CURRENT_SCRIPT_PATH%\sample-images\1.png
-)
-
-if "%CM_INPUT_2%" == "" (
-  set CM_INPUT_2=%CM_TMP_CURRENT_SCRIPT_PATH%\sample-images\2.png
-)
-
 echo Running author's code ...
 
 del /F /Q cm.png
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/1.png b/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/1.png
deleted file mode 100644
index 722f0944ee..0000000000
Binary files a/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/1.png and /dev/null differ
diff --git a/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/2.png b/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/2.png
deleted file mode 100644
index a681f89a5a..0000000000
Binary files a/cm-mlops/script/reproduce-ipol-paper-2022-439/sample-images/2.png and /dev/null differ
diff --git a/cm-mlops/script/reproduce-mlperf-inference-intel/customize.py b/cm-mlops/script/reproduce-mlperf-inference-intel/customize.py
index 7b1cf69a5c..05279e631e 100644
--- a/cm-mlops/script/reproduce-mlperf-inference-intel/customize.py
+++ b/cm-mlops/script/reproduce-mlperf-inference-intel/customize.py
@@ -36,7 +36,8 @@ def preprocess(i):
         pass
     elif "retinanet" in env.get('CM_MODEL'):
         pass
-
+    elif "gptj" in env.get('CM_MODEL'):
+        env['CHECKPOINT_DIR'] = env['GPTJ_CHECKPOINT_PATH']
 
     script_path = i['run_script_input']['path']
     if env['CM_MODEL'] == "retinanet":
@@ -63,18 +64,29 @@ def preprocess(i):
             env['CM_MLPERF_INFERENCE_INTEL_HARNESS_PATH'] = os.path.join(os.getcwd(), "harness", "build", "gptj_inference")
             env['DATA_PATH'] = os.path.join(os.getcwd(), "harness", "gptj")
             env['MLPERF_INFERENCE_ROOT'] = env['CM_MLPERF_INFERENCE_SOURCE']
-            env['CHECKPOINT_DIR'] = env['GPTJ_CHECKPOINT_PATH']
+            final_model_path = os.path.join(harness_root, "data", "gpt-j-int8-model")
+            env['INT8_MODEL_DIR'] = os.path.dirname(final_model_path)
+            env['CM_ML_MODEL_PATH'] = final_model_path
 
     elif env['CM_LOCAL_MLPERF_INFERENCE_INTEL_RUN_MODE'] == "run_harness":
-        env['MODEL_PATH'] = os.path.dirname(os.path.dirname(env['CM_MLPERF_INFERENCE_INTEL_HARNESS_PATH']))
-        env['DATASET_PATH'] = os.path.dirname(os.path.dirname(env['CM_MLPERF_INFERENCE_INTEL_HARNESS_PATH']))
-        env['CM_RUN_DIR'] = i['run_script_input']['path']
-        env['CM_RUN_CMD'] = "bash run_harness.sh " + ("--accuracy" if env['CM_MLPERF_LOADGEN_MODE'] == "accuracy" else "")
-
         if env.get('CM_MLPERF_LOADGEN_MODE', '') == "compliance":
          audit_path = env['CM_MLPERF_INFERENCE_AUDIT_PATH']
          shutil.copy(audit_path, env['CM_RUN_DIR'])
 
+        if 'bert' in env['CM_MODEL']:
+            env['MODEL_PATH'] = os.path.dirname(os.path.dirname(env['CM_MLPERF_INFERENCE_INTEL_HARNESS_PATH']))
+            env['DATASET_PATH'] = os.path.dirname(os.path.dirname(env['CM_MLPERF_INFERENCE_INTEL_HARNESS_PATH']))
+            env['CM_RUN_DIR'] = i['run_script_input']['path']
+            env['CM_RUN_CMD'] = "bash run_bert_harness.sh " + ("--accuracy" if env['CM_MLPERF_LOADGEN_MODE'] == "accuracy" else "")
+        elif "gptj" in env['CM_MODEL']:
+            if env['CM_MLPERF_LOADGEN_MODE'] == "accuracy":
+                env['LOADGEN_MODE'] = 'Accuracy'
+            else:
+                env['LOADGEN_MODE'] = 'Performance'
+            env['INT8_MODEL_DIR'] = env['CM_ML_MODEL_PATH']
+            env['CM_RUN_DIR'] = i['run_script_input']['path']
+            env['CM_RUN_CMD'] = "bash run_gptj_harness.sh "
+
     return {'return':0}
 
 def postprocess(i):
diff --git a/cm-mlops/script/reproduce-mlperf-inference-intel/run_harness.sh b/cm-mlops/script/reproduce-mlperf-inference-intel/run_bert_harness.sh
similarity index 100%
rename from cm-mlops/script/reproduce-mlperf-inference-intel/run_harness.sh
rename to cm-mlops/script/reproduce-mlperf-inference-intel/run_bert_harness.sh
diff --git a/cm-mlops/script/reproduce-mlperf-inference-intel/run_gptj_harness.sh b/cm-mlops/script/reproduce-mlperf-inference-intel/run_gptj_harness.sh
new file mode 100755
index 0000000000..2b2c733476
--- /dev/null
+++ b/cm-mlops/script/reproduce-mlperf-inference-intel/run_gptj_harness.sh
@@ -0,0 +1,51 @@
+#!/bin/bash
+export PATH=${CM_CONDA_BIN_PATH}:$PATH
+
+export KMP_BLOCKTIME=4
+export KMP_AFFINITY=granularity=fine,compact,1,0
+export LD_PRELOAD=${LD_PRELOAD}:${CONDA_PREFIX}/lib/libiomp5.so
+export LD_PRELOAD=${LD_PRELOAD}:${CONDA_PREFIX}/lib/libtcmalloc.so
+
+export num_physical_cores=`lscpu -b -p=Core,Socket | grep -v '^#' | sort -u | wc -l`
+num_numa=$(numactl --hardware|grep available|awk -F' ' '{ print $2 }')
+
+NUM_PROC=$num_numa
+CPUS_PER_PROC=$((num_physical_cores/num_numa))
+WORKERS_PER_PROC=1
+TOTAL_SAMPLE_COUNT=13368
+BATCH_SIZE=14
+TIMESTAMP=$(date +%m-%d-%H-%M)
+HOSTNAME=$(hostname)
+#OUTPUT_DIR=offline-output-${HOSTNAME}-batch-${BATCH_SIZE}-procs-${NUM_PROC}-ins-per-proc-${WORKERS_PER_PROC}-${TIMESTAMP}
+
+export WORKLOAD_DATA=${CM_HARNESS_CODE_ROOT}/data
+export VALIDATION_DATA_JSON=${WORKLOAD_DATA}/validation-data/cnn_dailymail_validation.json
+
+cd ${CM_HARNESS_CODE_ROOT}
+OUTPUT_DIR="${CM_MLPERF_OUTPUT_DIR}"
+
+USER_CONF="${CM_MLPERF_USER_CONF}"
+
+
+	#--mode Performance \
+cmd="python runner.py --workload-name gptj \
+	--scenario Offline \
+	--mode ${LOADGEN_MODE} \
+	--num-proc ${NUM_PROC} \
+	--cpus-per-proc ${CPUS_PER_PROC} \
+	--model-checkpoint-path ${CHECKPOINT_DIR} \
+	--warmup \
+	--dataset-path ${VALIDATION_DATA_JSON} \
+	--batch-size ${BATCH_SIZE} \
+	--mlperf-conf ${CM_MLPERF_CONF} \
+	--user-conf ${CM_MLPERF_USER_CONF} \
+	--precision int8 \
+	--pad-inputs \
+	--quantized-model ${INT8_MODEL_DIR}/best_model.pt \
+	--workers-per-proc ${WORKERS_PER_PROC} \
+	--total-sample-count ${TOTAL_SAMPLE_COUNT} \
+	--output-dir ${OUTPUT_DIR} \
+	2>&1 | tee ${OUTPUT_DIR}.log"
+
+echo "$cmd"
+eval "$cmd"
diff --git a/cm-mlops/script/run-all-mlperf-models/_cm.yaml b/cm-mlops/script/run-all-mlperf-models/_cm.yaml
index 6404e53543..a53b73a193 100644
--- a/cm-mlops/script/run-all-mlperf-models/_cm.yaml
+++ b/cm-mlops/script/run-all-mlperf-models/_cm.yaml
@@ -1,24 +1,18 @@
+uid: 8d3cd46f54464810
 alias: run-all-mlperf-models
+
 automation_alias: script
 automation_uid: 5b4e0237da074764
+
 cache: false
+
 category: MLPerf benchmark support
-deps: []
-docker_input_mapping: {}
-input_description: {}
-input_mapping: {}
-new_env_keys: []
-new_state_keys: []
-post_deps: []
-posthook_deps: []
-prehook_deps: []
-tags: [
-  "run",
-  "natively",
-  "all",
-  "mlperf-models"
-]
-uid: 8d3cd46f54464810
+
+tags:
+- run
+- natively
+- all
+- mlperf-models
 
 variations:
 
@@ -134,5 +128,3 @@ variations:
     env:
       CATEGORY: edge
       DIVISION: closed
-
-versions: {}
diff --git a/cm-mlops/script/run-all-mlperf-models/run-cpp-implementation.sh b/cm-mlops/script/run-all-mlperf-models/run-cpp-implementation.sh
index 694edf6d69..704abff2d7 100644
--- a/cm-mlops/script/run-all-mlperf-models/run-cpp-implementation.sh
+++ b/cm-mlops/script/run-all-mlperf-models/run-cpp-implementation.sh
@@ -88,13 +88,13 @@ ${POWER} \
 run "cm run script --tags=generate-run-cmds,inference,_find-performance \
 --model=resnet50 --implementation=cpp --device=cuda --backend=onnxruntime \
 --adr.compiler.tags=gcc \
---execution-mode=valid \
---category=edge --division=open --scenario=Offline  --quiet --test_query_count=100"
+--test_query_count=20000 \
+--category=edge --division=open --scenario=Offline  --quiet"
 
 run "cm run script --tags=generate-run-cmds,inference,_find-performance \
 --model=retinanet --implementation=cpp --device=cuda --backend=onnxruntime \
 --adr.compiler.tags=gcc \
---execution-mode=valid \
+--test_query_count=2000   \
 --category=edge --division=open --scenario=Offline  --quiet"
 
 
@@ -139,3 +139,25 @@ run "cm run script --tags=generate-run-cmds,inference,_submission \
 --skip_submission_generation=yes \
 ${POWER} \
 --results_dir=$HOME/results_dir"
+
+#multistream
+run "cm run script --tags=generate-run-cmds,inference,_submission \
+--scenario=Offline \
+--model=resnet50 --implementation=cpp --device=cuda --backend=onnxruntime \
+--scenario=MultiStream \
+--category=edge --division=$division  --quiet \
+--adr.compiler.tags=gcc \
+--execution-mode=valid \
+--skip_submission_generation=yes \
+${POWER} \
+--results_dir=$HOME/results_dir"
+
+run "cm run script --tags=generate-run-cmds,inference,_submission \
+--model=retinanet --implementation=cpp --device=cuda --backend=onnxruntime \
+--scenario=MultiStream \
+--category=edge --division=$division  --quiet \
+--adr.compiler.tags=gcc \
+--execution-mode=valid \
+--skip_submission_generation=yes \
+${POWER} \
+--results_dir=$HOME/results_dir"
diff --git a/cm-mlops/script/run-docker-container/customize.py b/cm-mlops/script/run-docker-container/customize.py
index beceaff111..f01f9d1eb9 100644
--- a/cm-mlops/script/run-docker-container/customize.py
+++ b/cm-mlops/script/run-docker-container/customize.py
@@ -197,4 +197,4 @@ def update_docker_info(env):
     docker_image_tag = env.get('CM_DOCKER_IMAGE_TAG', docker_image_base.replace(':','-').replace('_','') + "-latest")
     env['CM_DOCKER_IMAGE_TAG'] = docker_image_tag
 
-    return
\ No newline at end of file
+    return
diff --git a/cm-mlops/script/run-mlperf-inference-app/customize.py b/cm-mlops/script/run-mlperf-inference-app/customize.py
index fe26967e47..eb96904b14 100644
--- a/cm-mlops/script/run-mlperf-inference-app/customize.py
+++ b/cm-mlops/script/run-mlperf-inference-app/customize.py
@@ -210,9 +210,12 @@ def get_valid_scenarios(model, category, mlperf_version, mlperf_path):
     import submission_checker as checker
 
     if "dlrm-99" in model:
-      model= model.replace("dlrm-99", "dlrm-v2-99")
+      model = model.replace("dlrm-99", "dlrm-v2-99")
+    if "sdxl" in model:
+      model = "stable-diffusion-xl"
 
     config = checker.MODEL_CONFIG
+
     internal_model_name = config[mlperf_version]["model_mapping"].get(model, model)
 
     valid_scenarios = config[mlperf_version]["required-scenarios-"+category][internal_model_name]
diff --git a/cm-mlops/script/run-mlperf-power-server/_cm.json b/cm-mlops/script/run-mlperf-power-server/_cm.json
index 23b47e8ae2..5082ca3f05 100644
--- a/cm-mlops/script/run-mlperf-power-server/_cm.json
+++ b/cm-mlops/script/run-mlperf-power-server/_cm.json
@@ -52,6 +52,15 @@
       "skip_if_env": {
         "CM_HOST_OS_TYPE": "windows"
       }
+    },
+    {
+      "names": [
+        "win32"
+      ],
+      "tags": "get,generic-python-lib,_package.pypiwin32",
+      "enable_if_env": {
+        "CM_HOST_OS_TYPE": "windows"
+      }
     }
   ],
   "tags": [
@@ -63,5 +72,11 @@
     "server",
     "power-server"
   ],
-  "uid": "5bc68aaf389a40bd"
+  "uid": "5bc68aaf389a40bd",
+  "docker": {
+    "device": "/dev/usbtmc0",
+    "port_maps": [
+      "4950:4950"
+    ]
+  }
 }
diff --git a/cm-mlops/script/run-mlperf-power-server/customize.py b/cm-mlops/script/run-mlperf-power-server/customize.py
index 87638b5e58..65c7830420 100644
--- a/cm-mlops/script/run-mlperf-power-server/customize.py
+++ b/cm-mlops/script/run-mlperf-power-server/customize.py
@@ -20,11 +20,16 @@ def preprocess(i):
         config.write(configfile)
     print({section: dict(config[section]) for section in config.sections()})
 
-    cmd = 'python3 ' + env['CM_MLPERF_POWER_SOURCE'] + '/ptd_client_server/server.py -c power-server.conf'
+    if env['CM_HOST_OS_TYPE'] == "windows":
+        cmd_prefix = ""
+    else:
+        cmd_prefix = "sudo "
+
+    cmd =  env['CM_PYTHON_BIN_WITH_PATH'] + ' ' + os.path.join(env['CM_MLPERF_POWER_SOURCE'], 'ptd_client_server', 'server.py') +' -c power-server.conf'
     if env.get('CM_MLPERF_POWER_SERVER_USE_SCREEN', 'no') == 'yes':
-        cmd = 'sudo screen -d -m ' + cmd + ' '
+        cmd = cmd_prefix + ' screen -d -m ' + cmd + ' '
     else:
-        cmd = 'sudo ' + cmd
+        cmd = cmd_prefix + cmd
 
     env['RUN_CMD'] = cmd
 
diff --git a/cm-mlops/script/run-mlperf-power-server/run.bat b/cm-mlops/script/run-mlperf-power-server/run.bat
new file mode 100644
index 0000000000..d23f0addf5
--- /dev/null
+++ b/cm-mlops/script/run-mlperf-power-server/run.bat
@@ -0,0 +1,7 @@
+@echo off
+
+echo %RUN_CMD%
+
+%RUN_CMD%
+
+IF %ERRORLEVEL% NEQ 0 EXIT %ERRORLEVEL%
diff --git a/cm/README.md b/cm/README.md
index f9c82f5cf0..99e6e2ccf3 100644
--- a/cm/README.md
+++ b/cm/README.md
@@ -15,8 +15,10 @@ from [MLCommons projects](https://mlcommons.org) and [research papers](https://c
 in a unified way on any operating system with any software and hardware
 either natively or inside containers.
 
-Here are some most commonly used examples from the community:
-
+Here are a few most commonly used examples from the CM users 
+that should run in the same way on Linux, MacOS, Windows and other platforms
+(see [Gettings Started Guide](docs/getting-started.md) to understand 
+how they work and how to reuse them in your projects):
 
 ```bash
 pip install cmind
@@ -24,31 +26,44 @@ pip install cmind
 cm pull repo mlcommons@ck
 
 cm run script "python app image-classification onnx"
+cmr "python app image-classification onnx"
 
-cm run script "download file _wget" --url=https://cKnowledge.org/ai/data/computer_mouse.jpg --verify=no --env.CM_DOWNLOAD_CHECKSUM=45ae5c940233892c2f860efdf0b66e7e
+cmr "download file _wget" --url=https://cKnowledge.org/ai/data/computer_mouse.jpg --verify=no --env.CM_DOWNLOAD_CHECKSUM=45ae5c940233892c2f860efdf0b66e7e
+cmr "python app image-classification onnx" --input=computer_mouse.jpg
 
-cm run script "python app image-classification onnx" --input=computer_mouse.jpg
+cm find script "python app image-classification onnx"
 
-cm docker script "python app image-classification onnx" --input=computer_mouse.jpg
-cm docker script "python app image-classification onnx" --input=computer_mouse.jpg -j -docker_it
+cmr "get python" --version_min=3.8.0 --name=mlperf-experiments
+cmr "install python-venv" --version_max=3.10.11 --name=mlperf
 
-cm run script "get generic-python-lib _package.onnxruntime"
-cm run script "get coco dataset _val _2014"
-cm run script "get ml-model stable-diffusion"
-cm run script "get ml-model huggingface zoo _model-stub.alpindale/Llama-2-13b-ONNX" --model_filename=FP32/LlamaV2_13B_float32.onnx --skip_cache
+cmr "get ml-model stable-diffusion"
+cmr "get ml-model huggingface zoo _model-stub.alpindale/Llama-2-13b-ONNX" --model_filename=FP32/LlamaV2_13B_float32.onnx --skip_cache
+cmr "get coco dataset _val _2014"
 
 cm show cache
 cm show cache "get ml-model stable-diffusion"
 
-cm run script "run common mlperf inference" --implementation=nvidia --model=bert-99 --category=datacenter --division=closed
+cmr "get generic-python-lib _package.onnxruntime" --version_min=1.16.0
+cmr "python app image-classification onnx" --input=computer_mouse.jpg
+
+cm rm cache -f
+cmr "python app image-classification onnx" --input=computer_mouse.jpg --adr.onnxruntime.version_max=1.16.0
+
+
+cmr "python app image-classification onnx _cuda" --input=computer_mouse.jpg
+
+cmr "cm gui" --script="python app image-classification onnx"
+
+cm docker script "python app image-classification onnx" --input=computer_mouse.jpg
+cm docker script "python app image-classification onnx" --input=computer_mouse.jpg -j -docker_it
+
+cmr "run common mlperf inference" --implementation=nvidia --model=bert-99 --category=datacenter --division=closed
 cm find script "run common mlperf inference"
 
 cm pull repo ctuning@cm-reproduce-research-projects
 cmr "reproduce paper micro-2023 victima _install_deps"
 cmr "reproduce paper micro-2023 victima _run" 
 
-...
-
 ```
 
 ```python
diff --git a/cm/cmind/__init__.py b/cm/cmind/__init__.py
index 55d7315691..315daab681 100644
--- a/cm/cmind/__init__.py
+++ b/cm/cmind/__init__.py
@@ -1,4 +1,4 @@
-__version__ = "1.6.0"
+__version__ = "1.6.0.1"
 
 from cmind.core import access
 from cmind.core import error
diff --git a/cmr.yaml b/cmr.yaml
index 125a72fca5..b94693a602 100644
--- a/cmr.yaml
+++ b/cmr.yaml
@@ -2,4 +2,4 @@ alias: mlcommons@ck
 git: true
 prefix: cm-mlops
 uid: a4705959af8e447a
-version: 1.5.4
+version: 1.6.0
diff --git a/docs/README.md b/docs/README.md
index 91ffa1a468..31f3bda991 100644
--- a/docs/README.md
+++ b/docs/README.md
@@ -9,7 +9,7 @@ The goal is to provide a common, simple and human-readable interface to manage,
 and customize diverse AI/ML apps, benchmarks and research projects across continuously changing
 models, datasets, software and hardware from different vendors in a unified and automated way.
 
-* [Getting Started tutorial](getting-started.md)
+* [Getting Started Guide](getting-started.md)
 * [Introduction](introduction-cm.md)
 * [CM installation and customization](installation.md)
 * [Unified CLI and Python API](interface.md)
diff --git a/docs/installation.md b/docs/installation.md
index 7a6444bb69..c9504757cc 100644
--- a/docs/installation.md
+++ b/docs/installation.md
@@ -223,6 +223,13 @@ cm rm repo mlcommons@ck --all
 cm pull repo --url={URL of the fork of github.com/mlcommons/ck}
 ```
 
+If you want to use stable CM snapshots of reusable automation recipes (CM scripts), 
+you can download a stable repository from Zenodo:
+```bash
+cm rm repo mlcommons@ck --all
+cm pull repo --url=https://zenodo.org/records/10576423/files/cm-mlops-repo.zip
+```
+
 If you use CM scripts with Python outside containers, we suggest you to set up CM Python virtual
 environment as described [here](../cm-mlops/automation/script/README-extra.md#using-python-virtual-environments).
 
diff --git a/docs/tutorials/mlperf-inference-power-measurement.md b/docs/tutorials/mlperf-inference-power-measurement.md
index c45e19c1a8..661c179b08 100644
--- a/docs/tutorials/mlperf-inference-power-measurement.md
+++ b/docs/tutorials/mlperf-inference-power-measurement.md
@@ -6,8 +6,8 @@
 
 ## Requirements
 
-1. Power analyzer (any one [certified by SPEC PTDaemon](https://www.spec.org/power/docs/SPECpower-Device_List.html)). 
-   Yokogawa is the one which most submitters have submitted with and a new single-channel model like 310E can cost around 3000$.
+1. Power analyzer (anyone [certified by SPEC PTDaemon](https://www.spec.org/power/docs/SPECpower-Device_List.html)). 
+   Yokogawa is the one that most submitters have submitted with and a new single-channel model like 310E can cost around 3000$.
    The [MLCommons taskforce on automation and reproducibility ](../taksforce.md) 
    is also using the [Yokogawa 310E](https://tmi.yokogawa.com/solutions/products/power-analyzers/digital-power-meter-wt300e/) 
    to automate and simplify MLPerf submissions.
@@ -29,7 +29,7 @@ An adapter like [this](https://amzn.to/3Cl2TV5) can help avoid cutting the elect
 
 ![pages (14)](https://user-images.githubusercontent.com/4791823/210117283-82375460-5b3a-4e8a-bd85-9d33675a5843.png).
 
-The director machine runs the `server.py` script and loads a server process which communicates with the SPEC PTDaemon. 
+The director machine runs the `server.py` script and loads a server process that communicates with the SPEC PTDaemon. 
 When a client connects to it (using `client.py`), it in turn connects to the PTDaemon and initiates a measurement run. 
 Once the measurement ends, the power log files are transferred to the client. 
 
@@ -52,12 +52,14 @@ to install the MLCommons CM automation tool.
 
 If you are having GitHub access to [MLCommons power](https://github.com/mlcommons/power) repository,
 PTDaemon should be automatically installed using the following CM command:
+
+PS: The below command will ask for `sudo` permission on Linux and should be run with administrator privilege on Windows (to do NTP time sync).
 ```bash
 cm run script --tags=mlperf,power,server --device_type=49 --device_port=/dev/usbtmc0
 ```
 * ``--interface_flag="-U" and `--device_port=1` (can change as per the USB slot used for connecting) can be used on Windows for USB connection
-* `--device_type=49` corresponds to Yokogawa 310E and `ptd -h` should list the device_type for all supported devices. Location of `ptd` can be found using the below command
-* `--device_port=20` and `--interface_flag="-g" can be used to connect to GPIB interface (currently supported only on Windows) with serial address set to 20
+* `--device_type=49` corresponds to Yokogawa 310E and `ptd -h` should list the device_type for all supported devices. The location of `ptd` can be found using the below command
+* `--device_port=20` and `--interface_flag="-g" can be used to connect to GPIB interface (currently supported only on Windows) with the serial address set to 20
 ```bash
 cat `cm find cache --tags=get,spec,ptdaemon`/cm-cached-state.json
 ```